We have discovered a critical tumor suppressor function for Pten in stromal fibroblasts during mammary tumor progression. Consequences of Pten/Ets2 pathway disruption in stromal fibroblasts in the MMTV-ErbB2 mammary tumor model include massive remodeling of the EGM and its constituents, increased inflammation and increased angiogenesis. Importantly, the Pten disruption signature can be found in human tumor stroma and can predict patient outcome. In collaboration with Project 3, we have recently found that the Pten signature is enriched in breast tumor stromal subclasses associated with non-luminal breast cancer and poor patient outcome. Based on these findings, our overall hypothesis is that defining the tumor microenvironment (TME) pathways activated by Pten loss in stromal fibroblasts will uncover the mechanisms of co-evolution of mammary tumor and stroma, findings that can be translated directly to human breast cancer. The proposal has three Aims: AIM 1: Identify gene expression networks regulated by Pten in the tumor microenvironment. Specific hypothesis to be tested: The differential effect of stromal Pten loss on the ErbB2 model is due to the activation of a specific pathway(s) that complements ErbB2 function. Expected Outcome: Networks that control cross-talk between different cell compartments in the TME and that correlate with patient outcome will be identified. AIM 2: Determine mechanisms by which stromal Pten pathways contribute to tumor progression. Specific Hypothesis to be tested: miR-320 is one Pten effector that controls epithelial cell migration and growth, as well as endothelial migration and/or proliferation, coupling tumor invasiveness and angiogenesis. Expected Outcome: We will experimentally verify the critical Pten pathways in the TME, and uncover the mechanisms involved. Aim 3: Use experimentally verified Pten-dependent TME networks to identify potential biomarkers for predicting outcome and selecting treatment strategies for human breast cancer. Specific Hypothesis to be tested: Stromal biomarkers present in multiple tumor microenvironment cell compartments will be useful in stratifying breast cancer patients and predicting patient outcome. Expected Outcome: We will reveal the pathways in the mouse models that are most relevant to human breast cancer, and begin testing their role as potential biomarkers for breast cancer subclasses.